With driving down, there’s a growing backlash against unnecessary highways | Vox.com

Brad Plumer writes about peak travel, overly optimistic forecasts, and the policy consequences at vox.com in: With driving down, there’s a growing backlash against unnecessary highways. He cites us in the closing paragraph:

Another key point was that states (and the federal government) should probably spend less money on building new roads and more money on maintaining existing roads. Economists have argued that more states should probably be doing this anyway — this Brookings Institution paper by Matthew Kahn and David Levinson makes the most comprehensive case. But the argument gets even stronger if vehicle travel is going to grow more slowly or plateau in the future.

What’s clear is that it doesn’t make much sense for states to keep planning around the idea that driving will grow at 20th century rates forever. Those predictions have been utterly wrong for nearly a decade now — and sticking with them could mean wasting billions on unnecessary roads and highways.


Study examines link between road networks and economic development |CTS Catalyst

From CTS Catalyst: Study examines link between road networks and economic development

Researchers studying the relationship between the growth of road networks and regional development at the county level in Minnesota found a weakening connection between the two.

“The influence of road networks on employment isn’t as large as it used to be,” according to Michael Iacono, a research fellow with the Department of Civil, Environmental, and Geo- Engineering (CEGE). Iacono and David Levinson, the RP Braun/CTS Chair in Transportation, authored a paper with their findings earlier this year titled “Mutual Causation in Highway Construction and Economic Development.” Iacono also shared their findings during a presentation at the CTS annual research conference in May.

The researchers measured regional development by examining changes in data about population and employment. They used 20 years of road data from the Minnesota Department of Transportation (1988–2007) and employment data from Minnesota counties for the same period.

Part of the reason for the lack of connection between roads and development simply may be the maturity of the road network. In other words, existing roads facilitate most economic activity. “We’ve made most of the biggest, most productive investments already,” Iacono explained.

The researchers interpreted the findings as evidence of a weakening influence statewide of road networks (and transportation more generally) on the location of economic development. However, it’s possible that other location factors such as human capital levels, tax rates, and natural amenities have become just as, if not more, important than transportation network considerations.

“One of the implications of the findings,” Iacono said, “is that highway projects ought to be evaluated according to the benefits they provide to users, rather than hoped-for impacts on local employment or land development.”

Iacono added that since, from a statewide perspective, most of the major investments in road networks have been made, further changes to road networks likely won’t lead to significant changes in where people live and work. But he cautioned that these findings are too broad to be applied to specific development projects.

The study did find, however, mutual causality between population changes and the growth of local networks. Specifically, it found that the growth of the highway network tends to follow changes in population, and this result was statistically significant. Population growth also followed highway growth (they were positively correlated), but the result was not statistically significant.

“Recent state highway and multimodal long-range plans align well with the research results, recognizing the importance of preserving existing high-value critical connections in a state of good repair,” says John L. Wilson, economic policy analyst with MnDOT’s Office of Transportation System Management. “Current funding constraints should limit system capacity expansion to cases with evident benefits to the traveling public in the form of travel time savings, reduced operating costs, and improved safety.”

The study stems from a previous research effort for MnDOT.

Fixing it first, the White House is on-board

Matthew Kahn wrote: “http://www.slideee.com/slide/economic-analysis-of-transportation-investments-the-grow-america-act, The White House cites my work with David Levinson (see page 4). ”

More recent research has highlighted the importance of selecting investments wisely in key areas of the country on the basis of their economic contributions. This research has also emphasized the importance of maintaining existing assets in a good state of repair.

A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King
A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King

For a copy of the Levinson and Kahn Classic: Fix it First, Expand it Second, Reward it Third, see: 


The roads and bridges that make up our nation’s highway infrastructure are in disrepair as a result of insufficient maintenance—a maintenance deficit that increases travel times, damages vehicles, and can lead to accidents that cause injuries or even fatalities. This deficit is in part due to a prioritization of new projects over care for existing infrastructure and contributes to a higher-cost, lower-return system of investment. This paper proposes a reorganization of our national highway infrastructure priorities to “Fix It First, Expand It Second, and Reward It Third.” First, all revenues from the existing federal gasoline tax would be devoted to repair, maintain, rehabilitate, reconstruct, and enhance existing roads and bridges on the National Highway System. Second, funding for states to build new and expand existing roads would come from a newly created Federal Highway Bank, which would require benefit-cost analysis to demonstrate the efficacy of a new build. Third, new and expanded transportation infrastructure that meets or exceeds projected benefits would receive an interest rate subsidy from a Highway Performance Fund to be financed by net revenues from the Federal Highway Bank.

How much does driving your car cost, per minute?

Kevin Hartnett writes in the Boston Globe: “How much does driving your car cost, per minute?”  about my recent post.

It’s often argued that the only way Americans will drive less is if driving becomes more expensive, through something like higher taxes or rising oil prices. A pointed blog post this week by David Levinson, a transportation systems expert at the University of Minnesota, suggests an even more straightforward approach: Show people that driving already costs a lot more than they realize.

A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King
A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King

Levinson begins with the fact that in Minneapolis, the car sharing company car2go charges $.38 a minute. That seems like a lot if you think of the cost of operating your own car as merely the cost of gasoline—at current prices that’s about $.05-$.10 per minute. Of course, you’re also on the hook for the purchase price of your car, plus insurance, taxes, and maintenance. Levinson tallies those costs and estimates the true out-of-pocket cost of driving is around $.235 per minute. He goes one step further and figures that when you add in the costs of driving-related externalites like noise, pollution, and traffic—costs that are ultimately shouldered by society as a whole, rather than any one person directly—driving costs $.34 per minute.

At that rate, the decision to hop in the car starts to look a little diffrent. The “true cost” of a 25-minute drive to mall comes to around $8.50, and another $8.50 the other way. That’s $17 for a trip to Best Buy.

Levinson wonders how much less people would drive if they took this all-in view. We know that when the price of gasoline doubles, people tend to drive about 5 percent less. If suddenly you revealed that each driving trip has been costing almost six-times what we thought it did, Levinson estimates we’d all cut our driving by 29 percent, nearly overnight.

Or maybe we wouldn’t. For one, it’s probably not fair to assume people are completely oblivious to the non-fuel costs of driving—we might not think about our monthly car payments every time we take a trip to the store, but that cost certainly factors into our decision to own a car in the first place. And second, as Levinson allows, there’s some amount of driving we’d do almost regardless of how much it costs. This isn’t to say that higher fuel taxes wouldn’t push people toward mass transit (they almost certainly would). The lesson might be, as Levinson suggests, that “pay-as-you-go” driving would reduce our use of cars—but it might equally be that driving is deeply embedded in our culture and our daily routines, and Americans are willing to pay seemingly irrational amounts of money to keep it that way.

Kevin Hartnett is a writer in South Carolina. He can be reached at kshartnett18@gmail.com.

We don’t pay enough for transportation

In the United States, drivers (and parenthetically transit users … but this post is about drivers) don’t pay enough for transportation. As a result, drivers use too much and have misleading anchors about what prices “should” be.  When drivers  are shown and charged the actual cost of things, they are surprised, and not a little bit disgruntled.

An illustration: car2go vs. Private Costs of Auto Ownership

Why should my car2go ride be $0.38 per minute? [Note, rates vary by city.] I don’t pay $0.38 per minute to ride my own car, or transit, do I?

When driving my car, I pay out-of-pocket for gasoline. At today’s prices of about $3.50 per gallon (inclusive of taxes), and 35 miles per gallon, I am paying just $0.10 mile out-of-pocket. At 30 miles per hour (2 minutes per mile), I am paying $0.05 per minute (assuming no variation in fuel economy).

Much of the other $0.33 per minute is paying for what we perceive to be the fixed costs of vehicle ownership: the cost of the vehicle itself, insurance, maintenance, and repairs.

  • The cost of the vehicle, say $15,000 for a vehicle that runs 100,000 miles before depreciating to $0, is $0.15/mile or $0.075/minute.
  • Insurance might run $1000 per year or $0.10/mile ($0.05/minute). (Pay as you drive (or pay at the pump) insurance is a long discussed policy that has yet to be mainstream in the US. A version exists in Australia and some other places, including some US opt-ins.)
  • Vehicle taxes are about a quarter to half that (depending on where you are), so let’s say $0.01/minute. In some states these are dedicated to infrastructure, so we need to be careful to avoid double counting.
  • Repairs, oil, and maintenance probably have a similar running cost to insurance, less in the early years, more in the later year ($0.05/minute).

Adding that together is $0.185/minute.


That leave $0.165/minute in “out-of-pocket” costs car2go charges above what you would pay for an equivalent vehicle. Some of this is car2go operating expenses — load balancing or moving cars around so they will be near you, paying the cities for “free” on-street parking, having a nice app and GPS. Some of this might be because car2go vehicles are actually used less than many private cars, so the fixed costs have to be spread over fewer minutes. Some of this might be higher insurance than you would pay. Some of this is on-road assistance as needed. Some of this is operating profit, car2go has to break-even as a business or it will cease to exist.

One point is the average user of car2go drives less than the average owner of an equivalent car, thereby saving the outlay of $15,000 for ownership, $1000 per year for maintenance, $1000 per year for fuel, $1000 per  year for insurance, and $250 per year for vehicle tabs. Thus when they are willing to drive, they pay more per minute than the per minute basis for an owned car because they are paying for the option value of having a car when they want, but not when they don’t.

The second point is if the average owner of a car paid an additional $0.185 per minute on top of $0.05 per minute, they would drive less. For a 10 minute trip, they would be out-of-pocket an extra $1.85. For a 30 minute trip, they would pay an extra $3.55, which is about the out-of-pocket price of an express bus. We will discuss this more later in the post.

Thus far we have shown over-consumption of private vehicle ownership costs due to bundling these costs as fixed rather than variable.

Non-Private Costs


This does not yet fully include the full cost of infrastructure. Nationally direct user fees (gas tax and tolls) pay for about one-third the cost of all roads, the other two-thirds comes from general revenue (particularly property taxes at local jurisdictions, but also pseudo-user fees like motor vehicle taxes).  Since fuel taxes (the bulk of highway user fees) are $0.184/gallon at the national level and $0.285 in Minnesota, $0.47 in total, we would need to about triple it for user fees to pay for all of infrastructure costs (to $1.41/gallon or $0.04/mile or $0.02/minute, of which $0.0067/minute is already covered by existing gas taxes, meaning a new tax of $0.0133/minute should be levied to convert road infrastructure costs into a user charge).

This assumes infrastructure spending is the right amount in total, about which there is considerable argument. Clearly much infrastructure is in poor quality or insufficient, which increases vehicle repairs, crash rates, congestion, and future infrastructure costs.


Parking is usually “free” where I and most Americans live and work and shop, so this is not an out-of-pocket cost until we start charging for parking. Obviously there is a cost that is bundled into other real estate transactions, or is subsidized by the infrastructure provider in the case of free on-street parking. The value of this land in alternate uses depends on location, and in most but not all US places approaches zero.


Like parking, this also does not include externalities, which are also “free”.  Crashes are mostly internalized in insurance, but congestion and pollution and CO2 emissions and noise are not internalized. The estimates on these vary widely. I looked at this a long time ago, and I think the logic is still valid, though valuations have certainly increased. (A more recent peer reviewed summary is here).  This might be on the order of magnitude of $0.20 per mile or $0.10 per minute, though again varies hugely based on location and assumptions. I think I am being generous here (assuming a high value rather than a low value), but there will always be someone with a higher value.

We have not also included user time. Presumably drivers consider their own time already (though undoubtedly over-estimate the time spent driving). At an average wage of $20 per hour (it is probably a bit higher) this would be $0.33 per minute of labor foregone. Note this is roughly the same level as the full monetary costs of travel. In Benefit Cost Analysis, transport economists typically use half the wage rate.



Summarizing our ballpark out-of-pocket monetary costs per minute:

  • $0.05 fuel (currently paid, including gas taxes)
  • $0.075 vehicle ownership
  • $0.05 Repairs, oil, and maintenance  (converting to a per minute charge)
  • $0.05 insurance (converting to a per minute charge)
  • $0.0133 additional fuel tax for transportation infrastructure (replacing vehicle taxes and general revenue)
  • $0.10 externalities (excluding crash externalities – see insurance)


  • $0.34 Total

Reduction in Travel

How much less travel would there be if the costs of driving paid out-of-pocket on a per use basis?  Economists use the elasticity of demand with respect to price to estimate this. This tells us how much demand drops as prices increase. The short run elasticity of demand for driving (measured in vehicle miles traveled) with respect to the price of gas is about -0.05, meaning for every 100% increase in the price of gas, there is a 5% decrease in gasoline consumption (which correlates to driving in the short run, in the long run there is also a shift in vehicle fuel economy). So if we hold that to be true for all costs, going from $0.05 per minute to $0.34 per minute is 676% higher cost (a 576% increase), leads me to expect about a 29% reduction in fuel use (mileage) in the short run if people paid their roughly fixed costs plus infrastructure plus externalities of vehicle ownership as variable costs instead. Of course at the magnitude of shift, the elasticity values may no longer hold. In any case, this is no small matter. Certainly the direction is right, countries with much higher fuel taxes see much less driving in general.

Economic Responses

Income Effects

There would be a countervailing income effect, as people now had an income that was higher by the cost of the car-payment (say $1500 per year over 10 years), and $2250 in other vehicle  costs. That additional income effect would be spent con onsumption of lots of goods, not just travel by automobile. A fraction of it (maybe 20-30%) would go back to pay for additional transportation, though not necessarily more travel, perhaps just nicer travel in a better car (a Smart ForFour instead of a Smart ForTwo). Since wealthier people travel a bit more than less wealthy, there would be a small bit more travel, but probably a relatively de minimis amount.


Induced Demand

Further, if we did reduce congestion, we would expect at least some people to take advantage of that change and start traveling more than they otherwise would have (there would be induced demand from the lower travel times).

Time is money

As the adage goes, “time is money”, and if we were more directly aware of the cost of our travel, we would spend far less on it. This implies we over-consume travel compared to a system that charged users directly for their full costs. As we move towards more efficient and equitable transportation funding, using road pricing, and an economy with vehicles as a service (car sharing, ride sharing, cloud commuting) we should expect significantly less travel demand.

Consequentially infrastructure providers should supply  less transportation capacity in this policy environment than one where people could free ride and over-consume. Since infrastructure is long-lived, planning for a smaller network should begin now, with the aim to avoid irreversible investments made today  that will later be seen as unnecessary.


Road Pricing in the United States

A special issue of  Research in Transportation Economics — Volume 44, Pages 1-70 (June 2014) Road Pricing in the United States, edited by Mark Burris, just came out. We have a paper in here, but the others are interesting as well. This is behind a paywall, so if your University doesn’t subscribe, you can’t get it directly, but I am sure individual authors would be happy to send copies, and pre-prints may be online.

Does road pricing affect port freight activity: Recent evidence from the port of New York and New Jersey
David A. King, Cameron E. Gordon, Jonathan R. Peters

The feasibility of modernizing the Interstate highway system via toll finance
Robert W. Poole Jr.

HOT or not: Driver elasticity to price on the MnPASS HOT lanes
Michael Janson, David Levinson

Using vehicle value as a proxy for income: A case study on Atlanta’s I-85 HOT lane
Sara Khoeini, Randall Guensler

The impact of HOT lanes on carpools
Mark Burris, Negin Alemazkoor, Rob Benz, Nicholas S. Wood

Theory versus implementation in congestion-priced parking: An evaluation of SFpark, 2011-2012
Daniel G. Chatman, Michael Manville

A framework for determining road pricing revenue use and its welfare effects
Timothy F. Welch, Sabyasachee Mishra

It’s “only” 5 minutes, or Green Line Delay Monetized

Jim Walsh reports in the Strib

“According to timetables released before the line opened, a trip from Union Depot to Target Field was expected to take about 48-49 minutes. Metro Transit officials said last week that the westbound Green Line is averaging about 54 minutes, end to end and that the eastbound train is averaging about 53 minutes.”

If we take these numbers at face value, the train is 5 minutes late on average. (It is probably worse than this from a user perspective, because the times when it is late is when more riders are on the train to experience its lateness, when few riders are on-board in off-peak periods, it probably runs much closer to on-time).

He also reports 30,000 rides per day using the line. I don’t know the average length of trip, but let’s assume it is 1/2 the distance of the line. (This may be too long, but it off-sets the fact that more people experience the delay than the on-time conditions). Thus the average passenger trip would be delayed about 2.5 minutes.

There would be 75,000 person minutes of delay per day. There are 1440 minutes in a day, so about 1250 hours per day, or 52 person lives are lost to excess time on the train.

At a Value of Time of $15 (just as a point of information MnDOT now uses $16 for auto-value of travel time savings per person hour, but maybe transit users have a lower VOT because they don’t mind being delayed so much because they can do other things on the train) per hour, this is $18,750 per day or $6.8 Million per year.

Over 30 years, this is $205M without discounting. With discounting at 2% this is about $152M.

In short, this is not a small miss that we can just ignore (saying that it’s only 5 minutes and no one goes end to end anyway), and everything that can be done should be done to make the line go as fast as possible with a minimum of delay.

This does not even consider the lower operating costs to MetroTransit from less delay.

The Transportation Empowerment Act vs. Phasing in Devolution One EV at a Time

Reihan Salam at The Agenda discusses The Transportation Empowerment Act as a Model for Conservative Policymakers

In short the TEA would ” lower the federal gas tax while shifting virtually all responsibility for funding existing and new roads to state governments over five years”.

This is in contrast with current law, which would keep the federal gas tax the same (and thus decreasing in buying power), or proposals to raise the gas tax to maintain buying power in the face of declining fuel sales due both to fuel economy and declining vehicle travel. This problem will worsen with fleet electrification.

A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King
A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King

The vast majority of travel is within the same county, and thus certainly the same state (See The Hierarchy of Roads, the Locality of Traffic, and Governance for data from GPS from Minnesota), especially for big states in the western half of the US. Thus the problem is largely a state not national problem. Where there is a large share of interstate travel, states are fond of tolls (as in the northeast corridor) [See my dissertation: On Whom the Toll Falls for theory and Why States Toll for empirical evidence].

I am very empathetic with the idea of Subsidiarity, that we should deal with problems at the lowest reasonable scale of government.  This mismatch (or correspondence problem) of jurisdictional authority and the locale of the problem leads to many inefficiencies. Just as the federal government should not fix potholes on my local street, and my homeowners association should not have a nuclear policy, roads should be dealt with, and funded, closest to the user without incurring excess costs due to losing economies of scale. States should (and in many case did) raise their gas taxes, and further share that revenue with local levels of government (replacing local property taxes and other sources of general revenue), to fix today’s potholes and weak bridges. (Or perhaps there would only be one level of government operating and maintaining all levels of roads in states, which might be more efficient – it is what many other utilities do).

However, I am also empathetic with the idea that there is an existing source of revenue (the existing level of federal gas tax) on which there is consensus, which should not be thrown away so that 50 more difficult political fights can be had to achieve the same level of revenue. Most of the federal gas tax is returned to the states in proportion to the amount that was generated in those states, and while there are federal government rules and regulations and stipulations that add to the cost of doing business, most of those rules and regulations are well-intentioned.

The conclusion I have come to is we should keep the federal gas tax at the level it is at, dedicate it to specific national purposes (Maintaining and Rehabilitating the National Highway System – i.e. Fix-it-First) and allow it to fade away in importance over time. While of course it is technically possible to make this change in five years, I think it needlessly accelerates the process. (I am also aware of the Overton Window, and staking out a more extreme position helps move the dialog in that direction.)

There will be inevitable change in highway funding with electrification, and little would be lost waiting until EVs and HEVs are, say, 25% or 50% of the fleet, and the country is ready for some form of state-based mileage fee in lieu of gas taxes, administered with an emergent national standard so there don’t have to be 20 transponders in each car or 50 vignette stickers on your door.

In the end not all problems are federal problems and not all solutions are federal solutions. But there are real problems, and immediately lowering the federal gas tax exacerbates the short run problem in transportation. I am not convinced that this cold turkey strategy of phasing out gas taxes in 5 years (or at at least luke-warm turkey) outweighs the negative effects of federal funding, such as building a more capital intensive system than states themselves could justify if they bore all the costs, and raising costs all around.

See also: Fix It First  and Enterprising Roads